The Role of the NADPH Oxidase Complex,p38 MAPK, and Akt in Regulating HumanMonocyte/Macrophage SurvivalYijie Wang*, Mandy M. Zeigler*, Gregory K. Lam*, Melissa G. Hunter, Tim D. Eubank, Valery V. Khramtsov,Susheela Tridandapani, Chandan K. Sen, and Clay B. Marsh

Dorothy M. Davis Heart and Lung Research Institute, Department of Internal Medicine, and Department of Surgery, Ohio State University,Columbus, Ohio

M-CSF induces PI 3-kinase activation, resulting in reactive oxygenspecies (ROS) production. Previously, we reported that ROS mediatemacrophage colony-stimulating factor (M-CSF)–induced extracellu-lar regulated kinase (Erk) activation and monocyte survival. In thiswork, we hypothesized that M-CSF–stimulated ROS products modu-lated Akt1 and p38 activation. Furthermore, we sought to clarifythe source of these ROS and the role of ROS and Akt in monocyte/macrophage survival. Macrophages from p47phox-/- mice, lacking akey component of the NADPH oxidase complex required for ROSgeneration, had reduced cell survival and Akt1 and p38 mitogen-activated protein kinase (MAPK) phosphorylation compared withwild-type macrophages in response to M-CSF stimulation, but hadno difference in M-CSF–stimulated Erk. To understand how ROSaffected monocyte survival and signaling, we observed that NACand DPI decreased cell survival and Akt1 and p38 MAPK phosphory-lation. Using bone marrow–derived macrophages from mice ex-pressing constitutively activated Akt1 (Myr-Akt1) or transfectingMyr-Akt1 constructs into human peripheral monocytes, we con-cluded that Akt is a positive regulator of monocyte survival. More-over, the p38 MAPK inhibitor, SB203580, inhibited p38 activityand M-CSF–induced monocyte survival. These findings demonstratethat ROS generated from the NADPH oxidase complex contributeto monocyte/macrophage survival induced by M-CSF via regulationof Akt and p38 MAPK.

Keywords: Akt; macrophage/monocyte; p47phox; p38 MAP; ROS

Monocytes are produced in the bone marrow, and in the absenceof serum or survival factors they die via apoptosis in 24–48 h.To evade apoptosis, monocytes depend on activation by growthfactors, such as macrophage colony-stimulating factor (M-CSF).We reported that the PI 3-kinase inhibitor LY294002 suppressesthe survival of human monocytes and reduces Akt1 and extracel-lular regulated kinase (Erk) activation in response to M-CSF(1, 2). M-CSF-induced Erk activation is driven by the productionof endogenous reactive oxygen species (ROS) and appears tobe an important mediator of monocyte survival to M-CSF stimu-lation (2). However, the influence of M-CSF–induced ROS on

(Received in original form May 8, 2006 and in final form August 8, 2006 )

* These authors contributed equally to this article.

This work was supported by NIH grants HL63800, HL66108, HL67176, GM69589,American Lung Association Johnie Walker Murphy Career Investigator Award, andthe Kelly Clark Foundation; G.L. was a fellow of the Stanley J. Sarnoff Endowmentfor Cardiovascular Science during the time of his research at The Ohio StateUniversity.

Correspondence and requests for reprints should be addressed to Clay B. Marsh,Room 201, Dorothy M. Davis Heart and Lung Research Institute, Ohio StateUniversity Medical Center, Columbus, OH 43210. E-mail: Clay.Marsh@osumc.edu

Am J Respir Cell Mol Biol Vol 36. pp 68–77, 2007Originally Published in Press as DOI: 10.1165/rcmb.2006-0165OC on August 24, 2006Internet address: www.atsjournals.org

CLINICAL RELEVANCE

This study provides insight into mechanisms underlyingROS-regulated cellular survival in normal monocytes andmacrophages. It also provides supportive evidence forROS-related cancer and inflammatory diseases therapy.

Akt1 activity or other survival pathways in mononuclear phago-cytes remains unknown.

Some proinflammatory cytokines, such as TNF-�, IL-1�,TGF-�, M-CSF, platelet-derived growth factor (PDGF), epider-mal growth factor (EGF), and angiotensin II, induce cellularROS production (3). While there are several oxidant-generatingcomplexes in phagocytes, PI 3-kinase–initiated products can acti-vate the NADPH oxidase system, a multi-component complexassembled at the membrane. Recruitment and the assemblyof the cytosolic components, Rac2, p47phox, and p67phox to themembrane-bound gp91phox and p22phox, are PI 3-kinase dependent(4–6). Patients with autosomal recessive chronic granulomatousdisease (CGD) can harbor mutations in which no membranetranslocation of Rac2, p47phox, and p67phox occurs (7). The role ofthe NADPH oxidase system in Akt/protein kinase B activation issupported by studies in Rac2-deficient murine mast cells, whichhave reduced NADPH oxidase activity and ROS production aswell as reduced Akt activation compared with wild-type cells(8). Rac2 is found in hematopoietic cells and is critical in theNADPH oxidase function (9). Data in this study demonstratethat ROS produced by the NADPH oxidase regulates mast cellsurvival through Akt activation.

There are three isoforms of Akt: Akt1, Akt2, and Akt3. Whileit is not clear if these isoforms have redundant or independentactivity, Akt1 activity has been linked to survival events in bothtransformed and normal cells. To promote cellular survival, Akt isactivated. Akt recognizes PI (3,4,5) P3 and PI (3,4) P2 via pleckstrinhomology domains (see review in Ref. 10). Once membrane-localized, Akt is activated by phosphorylation on threonine-308by the enzyme PDK1, promoting autophosphorylation of Akton serine residue 473. Alternatively, some reports suggest thatthe serine 473 phosphorylation of Akt is mediated by PDK2/MapKK, PKC-�2, or integrin-linked kinase (ILK). For maximalactivation, tyrosine phosphorylation of Akt by Src family kinasesalso appears important (see review in Ref. 11).

We reported that ROS mediate M-CSF–induced Erk activa-tion and monocyte survival; however, the source of oxidant gen-eration remained to be defined. Erk is a member of the mitogen-activated protein kinases (MAPKs). MAPKs consist of at leastsix major subfamily members, of which Erk, c-jun NH2-terminalkinase (JNK), and p38 MAPK are characterized. MAPKs regu-late cell proliferation, differentiation, motility, and survival inresponse to a wide variety of stimuli, including growth factors

Wang, Zeigler, Lam, et al.: Endogenous ROS Influence Akt and p38 MAPK 69

and oxidative stress. The exact function of MAPKs on cellularsurvival and apoptosis are complex (6). p38 MAPK can promoteeither cellular survival or apoptosis (see review in Ref. 12). Forexample, IL-24–induced apoptosis and expression of growtharrest– and DNA damage (GADD)–inducible genes in melanomacells are dependent on p38 MAPK. Similarly, cardiomyocytesand fibroblasts derived from p38 MAPK-� knockout mice aremore resistant to apoptosis. In contrast, p38 MAPK activationprotects neuronal PC12 cells from TNF-�–induced apoptosisand enhances osteoblastic SaOS-2 cell growth and chondrocytesdifferentiation. Other investigators reported that p38 MAPKplay no role in cell survival, as reported in thymocytes derivedfrom mice lacking either MMK3 or MMK6, which are upstreamof p38 MAPK activation. Thus, it appears that cell type andstimulus have a powerful influence on the role of p38 MAPKon cell life or cell death.

Increased phosphorylation of p38 MAPK is linked to ROSgeneration in neuronal AF5 cells with stimulation of neurotrans-mitter N-methyl-D-aspartate (NMDA) (13). Phorbol myristateacetate (PMA)-treated mast cell (HMC-1) was shown to stimu-late IL-8 and TNF-� production in a p38 MAPK/NF-�B–dependent manner (14). Since a majority of the data examiningthe regulation of p38 MAPK activity by ROS production andp38 MAPK–mediated cell survival involve cultured cell lines,we evaluated whether ROS-mediated p38 MAPK activation con-tributed to the survival of primary human monocytes.

In this work, we evaluated the influence of M-CSF–stimulatedROS generation on Akt activity, p38 MAPK phosphorylation,and cell survival in primary human monocytes and murine mac-rophages. We found that ROS produced by M-CSF stimulationinduced cellular survival by activating Akt and p38 MAPK innormal human monocytes and macrophages.

MATERIALS AND METHODS

Materials

Endotoxin-free RPMI 1640 and PBS (� 10 pg/ml) were purchased fromBioWhittaker (Walkersville, MD). FBS was obtained from HycloneLaboratories (Logan, UT). Recombinant M-CSF was purchased fromR&D Systems (Minneapolis, MN). DPI, LY294002, SB203580, andSB202474 were obtained from Calbiochem (San Diego, CA). Antibod-ies for Western blot analysis were obtained from Santa Cruz Biotech(Santa Cruz, CA) or Cell Signaling (Beverly, MA). All other reagentswere purchased from Sigma (St. Louis, MO) unless indicated otherwise.

Purification of Peripheral Blood Monocytes

Monocytes (66 � 2.1% CD14�) were isolated as previously describedfrom buffy coats obtained from the American Red Cross or healthyvolunteers following informed consent. Briefly, whole blood was diluted1:1 with 1� PBS and layered over histopaque-1077. The mononuclearlayer was clumped in RPMI 1640 medium supplemented with 10% FBSat 4C for 1 h. Cells were layered over FBS for 20 min and monocyteswere then collected.

Monocytes were also isolated using the Monocyte Isolation Kitfrom Miltenyi Biotech (Auburn, CA) according to the manufacturer’sprotocol. Antibody cocktail and magnetic beads coupled to an anti-hapten monoclonal antibody were added to the peripheral blood mono-nuclear cells (PBMC). PBMC were placed through a magnetic columnto separate monocytes from other cell types (86 � 3.3% CD14�).

After isolation, monocytes were resuspended at 1–10 � 106 cells/mlin RPMI 1640 medium supplemented with 10% FBS and 10 g/mlpolymyxin B for 1 h. Then the nonadherent cells were removed andcells were washed with warm RPMI 1640 medium. The monocytes wereincubated in RPMI 1640 medium with 10 g/ml polymyxin B for anotherhour before being subjected to inhibitors or M-CSF stimulation.

Isolation and Culturing ofBone Marrow–Derived Macrophages

Femoral and tibia bone marrow were isolated from p47phox-/-, myristoy-lated (Myr)-Akt1 or wild-type littermate mice. p47phox-/- mice were kindly

provided by Dr. Steve Holland (NIH, Bethesda, MD). Constitutivelyactive Myr-Akt1 transgenic mice were described previously and pro-vided by Dr. Michael Ostrowski (The Ohio State University, Columbus,OH) (15).

The bone marrow progenitor cells were plated in RPMI supple-mented with 10% FBS, 1% PSA (penicillin G sodium, streptomycinsulfate, and amphotericin B), 10 g/ml of polymyxin B, and 20 ng/mlof M-CSF. Cells were cultured in 37C incubators with 5% CO2 for5 d with the addition of M-CSF each day. Differentiated bone marrow–derived macrophages (BMM) were serum-starved for 1 d at 37C beforebeing re-stimulated with M-CSF.

Electron Paramagnetic Resonance Spectroscopy

Human monocytes (1 � 106/condition) were isolated and incubatedwith specific inhibitors or the dimethylsulfoxide (DMSO) control, thenwere left unstimulated or stimulated with M-CSF (100 ng/ml) for anadditional 20 min in the presence of the spin trap DMPO (5,5-dimethyl-1-pyrroline-N-oxide). The EPR spectra were collected at the indicatedtime points on a Varian E-9 EPR spectrometer (Palo Alto, CA). Typicalinstrument settings were microwave power: 20 mW, and modulationamplitude: 0.8 G, as previously described (2).

Annexin V and Propidium Iodide Staining

Cell apoptosis was measured using an Annexin V–FITC apoptosis de-tection kit according to manufacturer’s protocol (BD PharMingen, SanDiego, CA). Briefly, human monocytes (1 � 106) or murine BMM(5 � 105) were removed from culture dish by Accutase (eBioscience,San Diego, CA) and stained with Annexin V–FITC and propidiumiodide (PI) according to instructions. Samples were analyzed by flowcytometry (FACSCalibur; BD PharMingen). Early stage of apoptosiswas defined as Annexin V–positive PI-negative staining, and late stageof apoptosis and necrosis were defined as Annexin V and PI double-positive staining. In this study, we combined Annexin V–positive stain-ing and Annexin V/PI double-positive staining in our statistics.

Immunoprecipitation and Immunoblotting

Monocytes (10 � 106 cell/condition) were isolated, stimulated, and lysedon ice for 15 min in 1� lysis buffer (Cell Signaling) containing proteaseinhibitor cocktail III (Calbiochem). Lysates were cleared of insolublematerial then protein concentration determined (BioRad, Hercules,CA). The samples were separated by SDS-PAGE, transferred to anitrocellulose membrane, probed with the indicated antibodies, anddetected by ECL (Amersham Biosciences, Piscataway, NJ).

Transfection of Primary Human Monocytes

Monocytes were transfected using the Amaxa Nucleofector system(Amaxa, Cologne, Germany) with expression vectors containing cDNAcorresponding to Myr-Akt1 (kindly provided by Dr. D. Stokoe, CancerResearch Institute, University of California). The monocytes (10 � 106)isolated from Buffy coats were resuspended in Monocyte Solutioncontaining 0.5 g of DNA and transfected with program Y-01, thenimmediately washed in X-VIVO 15 serum-free media (BioWhittaker,Walkersville, MD) and plated in 12-well plates.

In Vitro p38 MAPK Assays

In vitro p38 kinase activity was measured using the assay kit from CellSignaling Technology (Beverly, MA) by following the manufacturer’sprotocol. In this method, equal amount of protein was immnuoprecipi-tated with immobilized p38 MAPK (Thr180/Tyr182) antibody. Thekinase reaction was performed in the presence of ATF2 fusion proteinand cold ATP. Phosphorylation of ATF2 was measured by Westernblot using phospho-ATF2Thr71 antibody. Since SB203580 is a reversiblep38 MAPK inhibitor, 100 nM of SB203580 was added in the cell lysisbuffer and kinase buffer.

Subcellular Fractionation

Monocytes (20 � 10 (6) cells/condition) were transfected with cDNAand plated in 12-well plates. Cells were cultured at 37C for 16–18 hand cells were removed from the plate and then cell membrane andcytosolic fraction were separated using the method described previously(16).

70 AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL 36 2007

Statistical Analysis

Each experiment was performed in duplicate. All statistical analysis wasperformed using the mean � SEM from individual donors. Differencebetween means was analyzed using ANOVA with Fisher’s post hoctesting. P � 0.05 was defined as statistically significant in these studies.

RESULTS

DPI Reduces M-CSF–Induced ROS Production

Previously we observed that production of ROS by M-CSF–stimulated monocytes is blocked by the PI 3-kinase inhibitorwortmannin (2). We also observed that ROS regulated Erk1/2in M-CSF–stimulated human monocytes. Using electron spinresonance (ESR) with DMPO as a spin trap, which forms arelatively stable paramagnetic adduct when reacting with freeradicals (17), we measured endogenous ROS produced byM-CSF–stimulated monocytes. The flavoprotein inhibitor DPIreduced ROS produced by M-CSF–stimulated monocytes to lev-els seen in unstimulated cells (Figure 1). In contrast, the vehiclecontrol agent DMSO, a known antioxidant, slightly reduced theEPR signal in response to M-CSF. These data confirmed ourprevious observations that M-CSF promotes ROS productionin human monocytes (2) and suggested that flavoprotein-containing oxidases such as NADPH oxidase were responsible.

M-CSF–Mediated ROS Production EnhancesMonocyte Survival

Since DPI and NAC block M-CSF–induced Erk activity andblocking MEK/Erk activity inhibits M-CSF–induced monocytesurvival (2), we were interested in determining the role of ROSon monocyte survival. To determine whether inhibition of en-dogenous ROS reduced M-CSF–stimulated human cell survival,we stimulated human monocytes with M-CSF in the presenceof DPI, NAC or the vehicle control DMSO (for DPI) and testedthe cells for survival. The addition of DPI or NAC enhancedAnnexin V/PI expression in M-CSF–stimulated monocytes com-pared to monocytes stimulated with M-CSF alone or with vehicletreatment. Importantly, M-CSF–stimulated cells treated with ei-ther DPI or NAC had Annexin V/PI staining to the level ofapoptotic cells grown in the absence of M-CSF (Figure 2A).

Notably, even 35–40% of the M-CSF–stimulated cell popula-tion underwent apoptosis as defined by Annexin V/PI staining.This amount of cell death is attributed to purification procedureand mechanical removal from the culture dish and is similar tothat of other reports (18). As an alternative method of measuringcell apoptosis, DNA fragmentation studies confirmed that DPIand NAC promoted oligonucleosomal DNA fragmentation inM-CSF–stimulated monocytes (data not shown). These observa-

Figure 1. DPI inhibits endogenous ROS produced by M-CSF–stimulatedmonocytes. Human monocytes (1 � 106/condition) were incubatedwith DPI (10 M) or DMSO control for 1 h and the cells were leftunstimulated (NS) or stimulated with M-CSF (100 ng/ml) for an addi-tional 20 min in the presence of the spin trap DMPO. EPR signal indicat-ing ROS production was measured and recorded. Data shown are repre-sentative of three independent donors.

Figure 2. Inhibition of endogenous ROS by antioxidants reduces cellsurvival, phosphorylation of Akt and p38 MAPK in human monocytes.(A ) Monocytes (1 � 106/condition) were isolated and incubated withDPI (10 M), NAC (20 mM), or DMSO control for 1 h, then stimulatedwith (�) or without (�) M-CSF for an additional 24 h and stained withAnnexin V/PI for flow cytometry. Unstimulated monocytes were usedas the 100% Annexin V/PI control. Data are expressed as the mean �

SEM of six independent donors (*P � 0.01 compared with M-CSF alonecontrol). (B ) Monocytes (10 � 106/ condition) pretreated for 1 h withthe indicated inhibitors were either unstimulated (�) or stimulated (�)with M-CSF (100 ng/ml) for 6 h, then lysed. An equal amount of proteinwas subjected to Western blot analysis using antibodies recognizingpAktThr308 and pAktSer473 or phospho-p38Thr180/Tyr182. (C ) The membraneswere stripped and reprobed for Akt1 or p38 MAPK. Shown in B and Cis a representative blot from six independent donors.

tions demonstrate a critical role for ROS in mediating M-CSF–induced cell survival.

M-CSF–Mediated ROS Production Stimulate Akt andp38 MAPK Phosphorylation

Since monocyte survival is mediated in a PI 3-kinase–dependentmanner (1), we next investigated whether Akt/protein kinase Bactivation by M-CSF was reduced in monocytes treated with

Wang, Zeigler, Lam, et al.: Endogenous ROS Influence Akt and p38 MAPK 71

NAC or DPI. The PI 3-kinase inhibitor LY294002 was usedas an active inhibitor of Akt in M-CSF–stimulated cells. Aktactivation was assessed at 7 h after stimulation because this timepoint corresponds to the initial activation of caspase-3 in serum-starved monocytes (19), an event preceding DNA fragmentationin monocyte apoptosis. As predicted, NAC, DPI and LY294002all reduced the phosphorylation of Akt (pAkt) following M-CSFstimulation in monocytes (Figure 2B). This observation indicatedthat ROS functioned upstream of Akt.

Next, we wanted to investigate whether the inhibition of ROSproduction reduced the phosphorylation of p38 MAPK in M-CSF–stimulated monocytes. Phosphorylation of p38 MAPK wasassessed at 7 h after stimulation with M-CSF. Similar to whatwe reported for Erk activation, p38 MAPK phosphorylation wasattenuated by the presence of NAC and DPI in M-CSF–stimulated human monocytes (Figure 2C). Importantly, the PI3-kinase inhibitor LY294002 did not reduce p38 MAPK phos-phorylation (Figure 2C). These data demonstrated that reducedproduction or action of ROS suppressed the activity of Akt andp38 MAPK in M-CSF–stimulated human monocytes.

NADPH Oxidase Plays a Selective Role in M-CSF–Induced Aktand p38 Activity

To define the source of ROS influencing Akt, p38 MAPK, andErk activation in M-CSF–stimulated monocytes, bone marrowmacrophages were derived from p47phox-/- or wild-type mice.These mice lack NADPH oxidase function and have significantlyreduced ROS production (20). Similar to Rac2-null mice mastcells, BMM from p47phox-/- mice had reduced Akt1 phosphoryla-tion compared with wild-type littermates in response to M-CSFstimulation (Figure 3A). Interestingly, p38 MAPK phosphoryla-tion was also reduced after M-CSF stimulation (Figure 3B) inthe cells from p47phox-/- mice, which was seen in all mice tested.Consistent with our finding, other investigators found that p38MAPK kinase activation is attenuated in vascular smooth musclecells from p47phox-/- mice compared with the wild-type mice inresponse to thrombin-activated ROS generation (21). Takentogether, those observations indicate that ROS produced fromthe NADPH complex function upstream of p38 MAPK andAkt1.

Since we reported that M-CSF-induced ROS modulated Erkactivity (2), we examined the effect of p47phox deficiency onM-CSF–stimulated Erk activity in macrophages. Our data showedno significant change in M-CSF–induced phosphorylation of Erkin p47phox-/- murine macrophages compared with wild-type cells(Figure 3C), as confirmed by densitometry analyses (data notshown). These observations suggest that ROS produced by theNADPH oxidase complex did not regulate Erk in M-CSF–stimulated macrophages. These data support a more direct rolefor ROS produced from the NADPH oxidase complex mediatingAkt1 or p38 MAPK activation in M-CSF–stimulated mononu-clear phagocytes, perhaps due to the cell membrane location ofthis oxidant-producing complex.

ROS from the NADPH Complex Affect Cellular Viability

We next determined if reduced Akt1 activity in p47phox-/- murinemacrophages played a role in cellular survival. Notably, p47phox-deficient mice are not phenotypically different than wild-typemice. Analysis of peripheral blood demonstrated normal num-bers of circulating blood monocytes and other myeloid andlymphoid lineages in p47phox-/- and wild-type animals (data notshown), suggesting that NADPH oxidase function did not playa critical role in native cellular survival or proliferation. In con-trast, the number of BMM produced by incubating the cells withM-CSF was reduced in cells from p47phox-/- mice compared with

Figure 3. Macrophages from p47phox-/- mice have decreased Akt1 andp38 MAPK activation, but not Erk activation. BMM were derived fromp47phox-/- or wild-type littermate mice (WT) by growth in the presenceof M-CSF (20 ng/ml) for 5 d. The cells were serum-starved overnight,then stimulated with M-CSF (10, 50, or 100 ng/ml) for 5 min or leftunstimulated. (A ) Akt1 activation was assessed using equal amount ofprotein by Western blotting with antibodies to pAktThr308 (upper panel)and pAktSer473 (middle panel). Membranes were reblotted for total Akt1(lower panel). (B ) Western blot analysis using phospho-p38 MAPK anti-body and equal loading was confirmed by p38 MAPK antibody. (C )Erk phosphorylation was detected using pErk 42/44 antibody. Equalloading is shown using Erk2 antibody. Shown are representative datafrom macrophages obtained from six independent mice.

the wild-type mice at 0, 10, and 50 ng/ml of M-CSF (Figures 4Aand 4B). Interestingly, staining for Annexin V/PI demonstratedenhanced survival only at M-CSF doses of 50 and 100 ng/ml inp47phox-/- cells, suggesting that other pathways compensated forthe loss of these ROS at basal levels of M-CSF (Figure 4C).These data indicated the NADPH oxidase complex was involvedin cellular survival in M-CSF–stimulated macrophages.

Macrophages from Myr-Akt1 Mice Have Enhanced Survival

Since macrophages from the p47phox-/- mice had reduced Akt1activity and cell survival in response to M-CSF stimulation, wenext directly assessed the role of Akt1 in regulating mononuclearphagocyte survival. We examined macrophages from mice express-ing a Myr-Akt1 isoform expressed in mononuclear phagocytes

72 AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL 36 2007

Figure 4. Macrophages fromp47phox-/- mice have reducedsurvival in response to M-CSF.BMM were derived fromp47phox-/- or WT mice. (A ) Cul-tured BMM (5 � 105/condi-tion) were analyzed for totalcell number in response toM-CSF using an Olympus 1X50inverted microscope (Olympus,Center Valley, PA) equippedwith a �40 objective and Nikoncamera. (B ) Percent change intotal cell number of M-CSFtreated p47phox-/- and wild-typeBMM compared with unstimu-lated BMM (0). Cell numberswere significantly decreased inthe p47phox-/- mice BMM com-pared with wild-type BMM atM-CSF concentrations of 10 and50 ng/ml (*P � 0.05). Datafrom A and B represent SEM ofvalues obtained from macro-phages obtained from twop47phox-/- and two wild-typemice. (C ) BMM (5 � 105/con-dition) were collected andstained with Annexin V/PI.Data are expressed as the mean� SEM from macrophagesfrom nine independent mice(*P � 0.04 compared with BMMnot stimulated with M-CSF).

controlled by a c-Fms promoter. The myristoylation tag on Akt1targets the protein to the membrane in close proximity to thePDK1 enzyme, rendering it constitutively active. Bone marrowfrom Myr-Akt1 and wild-type mice was isolated, differentiated,and cultured in serum-free media to examine differences in cellsurvival and differentiation. As expected, in the absence of stimu-lation, Akt1 was constitutively phosphorylated in Myr-Akt1BMM, but not in wild-type BMM (Figure 5A). To assess theinvolvement of Akt in cell survival, BMM cells were differenti-ated over 5 d in M-CSF (20 ng/ml) and then serum-starved for3 d. The Myr-Akt1 macrophages were resistant to basal celldeath as demonstrated by less positive Annexin V/PI stainingcompared with wild-type cells (Figure 5B). As anticipated, after

serum deprivation there were more live cells from Myr-Akt1mice compared with cells from wild-type mice (Figures 5C and5D). Collectively, these observations demonstrate the impor-tance of Akt1 in regulating macrophage survival.

The Effect of Myr-Akt1 Expression on HumanMonocyte Survival

Since Myr-Akt1 expression in BMM increased cell survival inthe absence of M-CSF, we next wanted to directly assess therole of Akt1 in human mononuclear phagocyte survival. First,we examined whether active Akt1 promoted the survival ofhuman monocytes. To address this issue, we expressed Myr-Akt1 cDNA constructs in human monocytes. For these studies,

Wang, Zeigler, Lam, et al.: Endogenous ROS Influence Akt and p38 MAPK 73

Figure 5. Myr-Akt1–expressing macrophages haveprolonged survival. BMM were generated fromMyr-Akt1 or WT littermate mice by growth in M-CSF(20 ng/ml) for 5 d. The cells were serum starved for24 h and (A ) Akt1 activation was assessed by Westernblotting with antibodies recognizing pAktThr308 andpAktSer473. The membranes then were reprobed for totalAkt1. (B ) The cells were incubated in RPMI 1640 me-dium for an additional 2 d to evaluate cell survival.BMM (5 � 105/condition) were removed from the cul-ture plates by Accutase, then stained with Annexin V/PIand analyzed by flow cytometry. Data shown representmacrophages from six Myr-Akt and six wild-type mice(*P � 0.05 when comparing Myr-Akt1 cell numbersversus wild-type cells). (C ) The pictures were taken us-ing �40 objective from Olympus IX50 inverted micro-scope equipped with a digital camera. (D ) Cells werecounted and there was a significant increase in the Myr-Akt1 mice BMM (*P � 0.05 compared with the WTcontrol).

we used the nucleofector system, which directly delivers DNAto the nucleus. We optimized this procedure to minimize celldeath and maintain high efficient transfer of recombinant DNA.As shown in Figures 6A and 6B, primary human monocyteswere transfected with � 40% transfection efficiency using GFPplasmid control. Western blotting with phospho-Akt antibodyindicated that transfection of Myr-Akt1 resulted in the constitu-tive activation of Akt1 without any stimulation (Figure 6C, upperpanel). Expression of the transfected Myr-Akt1 in monocyteswas confirmed (Figure 6C, middle panel) as well as equal proteinloading with �-actin (Figure 6C, lower panel). Since the Myr tagon Akt1 targets the protein to the membrane, we separatedmembrane and cytosolic fractions and verified that Myr-Akt1was targeted to the cell membrane after transfection (Figure6D). To examine the role of the transfected Myr-Akt on survivalin primary human monocytes, we incubated the transfectedmonocytes in serum-free medium devoid of growth factors for7 h, stained the cells with Annexin V/PI, and analyzed them byflow cytometry. As shown in Figure 6E, Myr-Akt1–transfectedcells had lower levels of Annexin V/PI staining compared tomock-transfected cells (P � 0.05). These data agree with otherstudies showing that constitutively activated Akt1 is vital forthe survival of in vitro differentiated macrophages from humanmonocytes (22).

p38 MAPK Regulates Human Monocyte Survival

We next examined the role of p38 MAPK in M-CSF–stimulatedmonocyte survival. Annexin V/PI staining of cells was assessedat 7 and 24 h in the presence or absence of the p38 MAPKinhibitor SB203580 or its inactive form SB202474 after the addi-

tion of M-CSF. We did not observe significant changes in cellularsurvival when cells were treated with SB203580 for 7 h (datanot shown). However, doses of SB203580 (100 nM–10 M) de-creased M-CSF–regulated monocyte survival after 24 h of incu-bation (Figure 7A). Since several investigators reported thatSB203580 inhibited insulin- (23) and IL-2–induced Akt1 phos-phorylation (24), we next investigated whether SB203580 inter-fered with Akt1 activation. Not surprisingly, we found thatSB203580 at concentration of 100 nM did not affect Akt1 phos-phorylation, while concentrations exceeding 1 M of SB203580or addition of the PI 3-kinase inhibitor LY294002 as a controlabolished the phosphorylation of Akt in response to M-CSF(Figure 7B). We reasoned that higher dosage of SB203580 re-duced monocyte survival due to the specific inhibition of p38MAPK and nonspecific inhibition of Akt1, resulting in apoptosis.To show that SB203580 specifically inactivated p38 MAP kinaseactivity, we performed a kinase assay using the ATF2 fusionprotein as a substrate. As shown in Figure 7B, the active p38MAPK inhibitor SB203580, but not its inactive analogueSB202474, inhibited p38 MAP kinase activity across a dose rangeof the inhibitor (100 nM–10 M) (Figure 7B). These observationssuggest that Akt played an essential role in monocytes survivalin response to M-CSF.

DISCUSSION

ROS-mediated cell survival is dependent on the level and sourceof ROS, and the phase of the cell cycle (25). Here, we demon-strate the selective influence of NADPH oxidase complex onAkt1 and p38 MAPK activation and their subsequent role on

74 AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL 36 2007

Figure 6. Regulation ofAkt1 in primary humanmonocytes influencescellular survival. Mono-cytes (10 � 106/condi-tion) were isolated frombuffy coat and tran-siently transfected withcDNA expressing eGFPor Myr-Akt1 using theAmaxa nucleofection.After washing, cellswere plated in RPMImedium supplementedwith 10% FBS and poly-myxin B (10 g/ml) andincubated at 37C in a5% CO2 incubator for 7h. The monocytes wereanalyzed by (A ) flow cy-tometry for transfectionefficiency and (B ) pho-tography using Olym-pus 1X50 inverted fluo-rescence microscope.Images shown are phasecontrast, brightfield (leftpanel), and fluorescencealone (right panel). (C )Cells were lysed 7 h aftertransfection and sub-jected to Western blotanalysis with phospho-Akt1 antibody (upperpanel). The membranewas subsequently blot-ted with total Akt1 anti-body (middle panel)and then blotted withanti–�-actin antibody toensure equal loading(lower panel). (D ) Aftertransfection, membraneand cytosolic fraction ofthe cells were also ana-lyzed by Western blot-ting with anti-Akt1 anti-

body to show expression of Myr-Akt1 protein (upper panel). The same membranes were reprobed with the membrane-specific anti-� adaptinantibody to show the purity of the membrane fraction (lower panel). (E ) Cells were stained with Annexin V/PI and analyzed by flow cytometry.Data are expressed as the mean � SEM from six independent donors. (*P � 0.05 compared with the mock transfected samples).

cellular survival in M-CSF–stimulated mononuclear phagocytes.While Erk1/2 is at least partially regulated by ROS in M-CSF–stimulated monocytes and plays a role in monocyte survival(2), our data suggest that the NADPH oxidase complex is notthe source of ROS-stimulating Erk activation. ROS fromthe NADPH oxidase complex appeared to regulate Akt1 andp38 MAPK activation in M-CSF–stimulated mononuclearphagocytes.

It is possible that PI 3-kinase-associated products selectivelyregulate the NADPH oxidase complex and target Akt1 andp38 MAPK but not Erk1/2. It has been reported that the phoxhomology (PX) domain of p47phox binds PtdIns (3,4) P2 resultingin p47phox translocation to the cell membrane and the subsequentactivation of the NADPH oxidase system (26). Interestingly, invascular smooth muscle cells, p47phox associates with the actincytoskeleton (27). Disruption of this interaction is reported to

reduce angiotension II–induced ROS production and decreasethe phosphorylation of Akt and p38 MAPK, but does not affectthe activation of Erk1/2. This observation is consistent with ourfindings that in M-CSF–stimulated murine macrophages, Akt1and p38 MAPK, but not Erk, activation was reduced in theabsence of p47phox.

While our data showed that macrophages from p47phox-/- ani-mals had less phosphorylation of Akt1 and p38 MAPK andreduced cell survival compared with wild-type cells, the biologi-cal role of the NADPH oxidase complex in inflammatory dis-eases involving macrophages is not clear. Some reports showthat cross-breeding p47phox-/- with apolipoprotein E-/- mice resultsin less atherosclerosis than in cross-bred wild-type C57Bl/6 andapolipoprotein E-/- mice (28), whereas others find no effect (29).These discrepant findings may relate to the impact of theNADPH oxidase complex on cell survival signals in growth

Wang, Zeigler, Lam, et al.: Endogenous ROS Influence Akt and p38 MAPK 75

factor–stimulated monocytes and also may be selective to thespecific growth factors present in the environment. For example,in the presence of M-CSF, we surmise that ROS produced fromboth the NADPH oxidase complex, activating Akt1 and p38,and ROS from an alternate source, activating Erk, are neededto maximally facilitate monocyte survival. This hypothesis issupported by our current and previous findings (2) that bothNAC and DPI significantly reduced M-CSF–induced monocytesurvival and both Akt1 and Erk activation are reduced by theseagents.

Since NAC and DPI reduced M-CSF–stimulated humanmonocyte survival and p38 MAPK phosphorylation, we rea-

Figure 7. Inhibition of p38 MAPK activation reduces human monocytesurvival. (A) Monocytes (1 � 106/condition) were isolated and platedin RPMI 1640 medium and incubated with SB203580 or SB202474(100 nM–10 M), LY294002 (20 M), or DMSO control for 1 h andthen stimulated with 100 ng/ml of M-CSF and incubated for 24 h.The cells were then stained with Annexin V/PI and analyzed by flowcytometry. Data are expressed as the mean � SEM from three differentdonors (*P � 0.05 compared with M-CSF–stimulated cells). (B) Mono-cytes (10 � 106/condition) were incubated in RPMI 1640 mediumsupplemented with polymyxin B (10 g/ml), M-CSF (20 ng/ml) over-night. The cells were serum starved for 2 h and incubated withSB203580, SB202474, LY294002, or DMSO control for 1 h beforestimulating with 100 ng/ml of M-CSF for 5 min. Whole cell lysates wereimmunoblotted with anti–phospho-Akt1 antibody (upper panel). Themembrane was reblotted with total Akt1 antibody (middle panel). p38kinase assay was performed with ATF2 fusion protein (lower panel). Datashown are representative of six independent donors.

soned that p38 activity was also important in M-CSF–inducedcell survival. Using human monocytes, we observed that thep38 MAPK inhibitor SB203580 at the concentration of 100 nMefficiently inactivated p38 kinase activity, but did not reduceAkt1 phosphorylation. Importantly, this dose of the p38 inhibitoralso reduced M-CSF–stimulated cell survival after 24 h of incuba-tion. Collectively our data indicated that Akt1 and p38 MAPKare positive regulators of monocyte survival in response toM-CSF stimulation. Figure 8 illustrates the proposed pathwaythat Akt and p38 MAPK pathways play in ROS and M-CSF–induced cellular survival.

Zhu and colleagues reported that NAC blocks G-CSF–induced Akt phosphorylation but not Erk1/2 or p38 MAPKphosphorylation in hematopoietic cell lines (5). However, in ourstudies using primary cells (2), we found that ROS inhibitorsNAC and DPI blocked M-CSF–induced phosphorylation of Akt,Erk1/2, and p38 MAPK. Therefore, cellular response to ROSappears to differ between primary cells and cell lines. Further-more, we found that both Akt1 and p38 MAPK phosphorylationwere decreased in BMM from p47phox-/- mice and, therefore ROSgeneration appears to be upsteam of Akt1 and p38 MAPK inM-CSF–induced cellular signaling. In accordance with our data,Gao and coworkers proposed that ROS are upstream of Akt inleukemia cells (30). Although specific targets regulated by ROSin M-CSF–stimulated human monocytes are not clear, proteinscontaining redox-sensitive cysteine or methionine groups aresusceptible to oxidation. Proteins important in cellular homeo-stasis fall into this category and include caspases, phosphatases,and transcription factors. Signaling molecules important in cellu-lar survival, such as Akt1, caspase-3, and transcription factorsincluding NF-�B, are modifiable by reactive oxygen or reactivenitrogen species (see review in Ref. 31).

Figure 8. Proposed simplified model for the NADPH oxidase systemand M-CSF–induced human monocyte/macrophage survival involvingAkt and p38 MAPK pathways.

76 AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL 36 2007

It has been shown that cancer cells and normal cells differin their ability to respond to ROS (32). In this article, we providedevidence that M-CSF–induced ROS production cooperates withAkt and p38 MAPK to promote monocyte and macrophagesurvival. Similarly, low concentrations of H2O2 accelerate woundhealing in p47phox-/- mice by facilitating wound angiogenesisin vivo (33). In contrast, Mochizuki and colleagues reported thatdepletion of ROS by DPI or siNox4RNAs to inhibit NADPHoxidase 4 from human pancreatic adenocarcinoma cell line in-duces apoptosis and reduces Akt-Ask1 activity (34). Studies havedemonstrated that increasing ROS production and downregulat-ing Akt through chemotherapeutic drugs are effective to induceapoptosis in leukemia cells but not normal bone marrow cells(30, 35).

In summary, we provide several lines of evidence supportingan important role for the NADPH oxidase complex in influenc-ing Akt1 and p38 MAPK activation. Furthermore, we show thatAkt1 and p38 MAPK are important in monocyte/macrophagesurvival regulated by M-CSF. Understanding the influence ofNADPH oxidase complex and ROS on monocyte survival andthe function of Akt1 and p38 MAPK activation may providenew opportunities to influence cellular inflammation and cancertherapy.

Conflict of Interest Statement : None of the authors has a financial relationshipwith a commercial entity that has an interest in the subject of this manuscript.

Acknowledgments : The authors thank Drs. Chris Baran and Judy Opalek for theirinsightful discussions and technical assistance. The authors are grateful for theassistance provided by the Flow Cytometry core facility at the HLRI of the OhioState University.

References

1. Kelley TW, Graham MM, Doseff AI, Pomerantz RW, Lau SM, OstrowskiMC, Franke TF, Marsh CB. Macrophage colony-stimulating factorpromotes cell survival through Akt/protein kinase B. J Biol Chem1999;274:26393–26398.

2. Bhatt NY, Kelley TW, Khramtsov VV, Wang Y, Lam GK, Clanton TL,Marsh CB. Macrophage-colony-stimulating factor-induced activationof extracellular-regulated kinase involves phosphatidylinositol 3-kinase and reactive oxygen species in human monocytes. J Immunol2002;169:6427–6434.

3. Baran CP, Zeigler MM, Tridandapani S, Marsh CB. The role of ROSand RNS in regulating life and death of blood monocytes. Curr PharmDes 2004;10:855–866.

4. Bokoch GM, Knaus UG. NADPH oxidases: not just for leukocytesanymore! Trends Biochem Sci 2003;28:502–508.

5. Zhu QS, Xia L, Mills GB, Lowell CA, Touw IP, Corey SJ. G-CSF inducedreactive oxygen species involves Lyn-PI3-kinase-Akt and contributesto myeloid cell growth. Blood 2006;107:1847–1856.

6. Touyz RM, Chen X, Tabet F, Yao G, He G, Quinn MT, Pagano PJ,Schiffrin EL. Expression of a functionally active gp91phox-containingneutrophil-type NAD(P)H oxidase in smooth muscle cells from humanresistance arteries: regulation by angiotensin II. Circ Res 2002;90:1205–1213.

7. Noack D, Rae J, Cross AR, Ellis BA, Newburger PE, Curnutte JT,Heyworth PG. Autosomal recessive chronic granulomatous diseasecaused by defects in NCF-1, the gene encoding the phagocyte p47-phox: mutations not arising in the NCF-1 pseudogenes. Blood 2001;97:305–311.

8. Yang FC, Kapur R, King AJ, Tao W, Kim C, Borneo J, Breese R,Marshall M, Dinauer MC, Williams DA. Rac2 stimulates Akt activa-tion affecting BAD/Bcl-XL expression while mediating survival andactin function in primary mast cells. Immunity 2000;12:557–568.

9. Diebold BA, Bokoch GM. Molecular basis for Rac2 regulation of phago-cyte NADPH oxidase. Nat Immunol 2001;2:211–215.

10. Franke TF, Hornik CP, Segev L, Shostak GA, Sugimoto C. PI3K/Aktand apoptosis: size matters. Oncogene 2003;22:8983–8998.

11. Song G, Ouyang G, Bao S. The activation of Akt/PKB signaling pathwayand cell survival. J Cell Mol Med 2005;9:59–71.

12. Roux PP, Blenis J. ERK and p38 MAPK-activated protein kinases: afamily of protein kinases with diverse biological functions. MicrobiolMol Biol Rev 2004;68:320–344.

13. Yosimichi G, Nakanishi T, Nishida T, Hattori T, Takano-Yamamoto T,Takigawa M. CTGF/Hcs24 induces chondrocyte differentiationthrough a p38 mitogen-activated protein kinase (p38MAPK), and pro-liferation through a p44/42 MAPK/extracellular-signal regulatedkinase (ERK). Eur J Biochem 2001;268:6058–6065.

14. Kim JY, Ro JY. Signal pathway of cytokines produced by reactive oxygenspecies generated from phorbol myristate acetate-stimulated HMC-1cells. Scand J Immunol 2005;62:25–35.

15. Ganesan LP, Wei G, Pengal RA, Moldovan L, Moldovan N, OstrowskiMC, Tridandapani S. The serine/threonine kinase Akt Promotes Fcgamma receptor-mediated phagocytosis in murine macrophagesthrough the activation of p70S6 kinase. J Biol Chem 2004;279:54416–54425.

16. Wang Y, Keogh RJ, Hunter MG, Mitchell CA, Frey RS, Javaid K, MalikAB, Schurmans S, Tridandapani S, Marsh CB. SHIP2 is recruited tothe cell membrane upon macrophage colony-stimulating factor(M-CSF) stimulation and regulates M-CSF-induced signaling. J Immu-nol 2004;173:6820–6830.

17. Roy S, Khanna S, Wallace WA, Lappalainen J, Rink C, Cardounel AJ,Zweier JL, Sen CK. Characterization of perceived hyperoxia in iso-lated primary cardiac fibroblasts and in the reoxygenated heart. J BiolChem 2003;278:47129–47135.

18. van Engeland M, Ramaekers FC, Schutte B, Reutelingsperger CP. Anovel assay to measure loss of plasma membrane asymmetry duringapoptosis of adherent cells in culture. Cytometry 1996;24:131–139.

19. Fahy RJ, Doseff AI, Wewers MD. Spontaneous human monocyte apopto-sis utilizes a caspase-3-dependent pathway that is blocked by endotoxinand is independent of caspase-1. J Immunol 1999;163:1755–1762.

20. Nowicki PT, Flavahan S, Hassanain H, Mitra S, Holland S, Goldschmidt-Clermont PJ, Flavahan NA. Redox signaling of the arteriolar myogenicresponse. Circ Res 2001;89:114–116.

21. Brandes RP, Miller FJ, Beer S, Haendeler J, Hoffmann J, Ha T, HollandSM, Gorlach A, Busse R. The vascular NADPH oxidase subunitp47phox is involved in redox-mediated gene expression. Free RadicBiol Med 2002;32:1116–1122.

22. Liu H, Perlman H, Pagliari LJ, Pope RM. Constitutively activated Akt-1 isvital for the survival of human monocyte-differentiated macrophages.Role of Mcl-1, independent of nuclear factor (NF)-kappaB, Bad, orcaspase activation. J Exp Med 2001;194:113–126.

23. Talukdar I, Szeszel-Fedorowicz W, Salati LM. Arachidonic acid inhibitsthe insulin induction of glucose-6-phosphate dehydrogenase via p38MAP kinase. J Biol Chem 2005;280:40660–40667.

24. Lali FV, Hunt AE, Turner SJ, Foxwell BM. The pyridinyl imidazoleinhibitor SB203580 blocks phosphoinositide-dependent protein kinaseactivity, protein kinase B phosphorylation, and retinoblastoma hyper-phosphorylation in interleukin-2-stimulated T cells independently ofp38 mitogen-activated protein kinase. J Biol Chem 2000;275:7395–7402.

25. Hoidal JR. Reactive oxygen species and cell signaling. Am J Respir CellMol Biol 2001;25:661–663.

26. Kanai F, Liu H, Field SJ, Akbary H, Matsuo T, Brown GE, Cantley LC,Yaffe MB. The PX domains of p47phox and p40phox bind to lipidproducts of PI(3)K. Nat Cell Biol 2001;3:675–678.

27. Touyz RM, Yao G, Quinn MT, Pagano PJ, Schiffrin EL. p47phox associ-ates with the cytoskeleton through cortactin in human vascular smoothmuscle cells: role in NAD(P)H oxidase regulation by angiotensin II.Arterioscler Thromb Vasc Biol 2005;25:512–518.

28. Yokoyama M, Inoue N, Kawashima S. Role of the vascular NADH/NADPH oxidase system in atherosclerosis. Ann N Y Acad Sci 2000;902:241–247.

29. Kirk EA, Dinauer MC, Rosen H, Chait A, Heinecke JW, LeBoeuf RC.Impaired superoxide production due to a deficiency in phagocyteNADPH oxidase fails to inhibit atherosclerosis in mice. ArteriosclerThromb Vasc Biol 2000;20:1529–1535.

30. Gao N, Rahmani M, Shi X, Dent P, Grant S. Synergistic antileukemicinteractions between 2-medroxyestradiol (2-ME) and histone deacety-lase inhibitors involve Akt down-regulation and oxidative stress. Blood2006;107:241–249.

31. Irani K. Oxidant signaling in vascular cell growth, death, and survival: areview of the roles of reactive oxygen species in smooth muscle andendothelial cell mitogenic and apoptotic signaling. Circ Res 2000;87:179–183.

Wang, Zeigler, Lam, et al.: Endogenous ROS Influence Akt and p38 MAPK 77

32. Ungerstedt JS, Sowa Y, Xu WS, Shao Y, Dokmanovic M, Perez G,Ngo L, Holmgren A, Jiang X, Marks PA. Role of thioredoxin inthe response of normal and transformed cells to histone deacetylaseinhibitors. Proc Natl Acad Sci USA 2005;102:673–678.

33. Roy S, Khanna S, Nallu K, Hunt TK, Sen CK. Dermal wound healingis subject to redox control. Mol Ther 2006;13:211–220.

34. Mochizuki T, Furuta S, Mitsushita J, Shang WH, Ito M, Yokoo Y,Yamaura M, Ishizone S, Nakayama J, Konagai A, et al. Inhibition of

NADPH oxidase 4 activates apoptosis via the AKT/apoptosis signal-regulating kinase 1 pathway in pancreatic cancer PANC-1 cells.Oncogene 2006;25:3699–3707.

35. Dasmahapatra G, Rahmani M, Dent P, Grant S. The tyrphostinadaphostin interacts synergistically with proteasome inhibitorsto induce apoptosis in human leukemia cells through a reactiveoxygen species (ROS)-dependent mechanism. Blood 2006;107:232–240.